One of the most well known communication and social deficits exhibited by children and adults with autism is the inability to mentalize or make inferences about what others know, believe, think, and feel, which renders them to operate socially in the "blind" (Baron-Cohen), 1995, Tager-Flusberg, 2001). Although researchers have attempted to discover which specific deficit or deficits cause this "mindblindness", its basis remains in open question. The current proposal focuses on the issue by investigating whether deficits in theory of mind in young children with autism stem from an inability to form concepts for animates an inanimates. The main hypothesis of the proposed studies is that young children with autism may have difficulties in encoding goal-directed action, which is generally considered a crucial precursor to theory of mind. To examine this question, the proposal uses the inductive generalization of generalized imitation procedure with typically developing infants and young children ad well as young children with autism. One study will examine whether young children with autism understand that animals and not vehicles are goal-directed. A second and third study will examine on what basis young children with autism and typically developing infants and young children generalize goal-directed behavior to animals and not vehicles (e.g., surface features such as eyes or legs versus category membership). The findings will have relevance to cognition, development, clinical psychology, and neuroscience, and they will have potential implication for early diagnosis of autism as well as for targeting specific deficits in intervention with the goal of improving social functioning.